Jim Zemlin at OSCON: The Mysterious Work of the Linux Foundation

By James Turner
July 22, 2008 | Comments: 1

Or, you may download the file.

James Turner: James Turner of O'Reilly News; I'm speaking today with Jim Zemlin, who is the Executive Director of the Linux Foundation. Thank you for taking the time to speak to me.

Jim Zemlin: Thanks for having me.

JT: So a lot of people probably have heard about the Linux Foundation but may not know what it, it's kind of the trilateral commission of Open Source. It's heard of but no one really understands what it does.

JZ: Yeah.

JT: Why don't you talk a little bit about what the Linux Foundation is about?

JZ: We are the free masons; we're a grand conspiracy that is behind everything. In fact it's interesting you mention that because in many ways Linux is like that. Whenever I tell people what I do, I tell them you may never have heard of Linux. I'll be at a party with my wife or out somewhere and I'll say it to someone--I work at the Linux Foundation. And they'll go Linux; I think I've heard of that. But when you put it in the perspective of just how prolific Linux is it's easier to understand kind of not only what Linux is but what we do as an organization. When you record something on your DVR you're using Linux. When you make a trade on the New York Stock Exchange you're using Linux. When you listen to music on your digital media system you're using Linux. When you make a call on a Motorola cell phone you're using Linux. It's an operating system that is literally everywhere. It's in the Mars Rover; it's in embedded systems that control the heating and cooling in your office building. It's literally everywhere, just like the trilateral commission. So what we do in that world of this sort of mass development of computing platform networks across embedded systems, mobile devices, servers, desktops--our organization serves as a place that helps the industry work together to make sure that the development process itself is healthy. We do that by sponsoring the work of Linus Torvalds who works at the Linux Foundation. We support the kernel community in their development process. We provide a legal defense fund for the platform so that if anyone had a bone to pick with Linux we would step up to defend users of the Linux platform. We create standards that make Linux more efficient and allow it to build a more cohesive eco-system, so we build application standards that allow applications written for one version of Linux to work on a different version of Linux. We coordinate the development of various components off the Linux platform, Linux printing components, all the unsung parts of the platform that make it work. We work on a lot of that, so--. It's not the most glamorous work; it's not the stuff that grabs the spotlight, but it's stuff that really impacts the platform in its growth.

JT: Now it's fairly remarkable in that although if you go up to the distribution level there's a lot of diversity in distributions--everything from mobile phone distros where you wouldn't even know there was a Linux in there to things like Ubuntu and Ximian that are very user-friendly desktop environments. But it's fairly remarkable that they're all still running on one kernel, especially since it is an Open Source kernel and anyone could have forked it any time they wanted to.

JZ: Yeah; it--

JT: What do you kind of credit with keeping that consistency?

JZ: So what I credit is one, the remarkable innovation that the kernel community itself, it does; we just did a study on who writes Linux and the number of people participating in the kernel community has doubled in just the last few years. The reason that people all use the same kernel is because there's tremendous advantage; there's advantage from an efficiency point of view in terms of you don't want to maintain your own operating system over time, but there's also a tremendous advantage from an innovation point of view of collectively participating in the development of the kernel even though what you make isn't necessarily similar. A mobile device handset is very different than a server that runs the New York Stock Exchange. But they both benefit from participating in the kernel development process because a developer who is working for, let's say Motorola, and contributing code to the Linux kernel to allow for longer battery life in their cell phone is creating code that can then be reused in a Wall Street datacenter to reduce their power consumption, which is now becoming the majority cost of their computing. Power now represents the largest percentage of the ongoing cost of running these data centers and there's really no other parallel in computing that allows this unexpected innovation to be shared across completely different types of computing. So that's I think why you're seeing people stick with using the same kernel.

JT: Linux, in the really good old days, was a bunch of individual people contributing; a lot of the Linux development work today comes out of IBM and Novell and Red Hat and major players in the industry. Is it harder to kind of make them all play nice together than it was when it was, you might say, a more altruistic group of people?

JZ: It's interesting; I would take a little bit subtler view of that in that it is like the good old days in that there are key individuals who create the software across all of the projects. There are maintainers of each of these projects who are--who work for companies like IBM and Red Hat or Novell or Google or whoever but are also individuals in their own right who have trusted relationship with other individuals who might happen to work at a large corporation. And while many of these people are employed by those companies to work on specific technical issues that affect that company's goals relative to Linux, the people who are sort of the ombudsmen, these trusted individuals who have this network of trusted developer relationships sort of act as a go-between to buffer the sort of often competitive corporate interests from the need to get along in order to get stuff done in the kernel development process. So it's interesting to see that this morphing of corporate interests and the competitiveness and sometimes fragmentation that comes from that are those different corporate goals being tempered by these tremendous individuals who work within the Open Source community.

JT: The Linux kernel has stayed very consistent but when you get above that level there is a lot of diversity and some people might say, and I might say occasionally, sometimes needless diversity with multiple desktops and nine different sound systems, and do you think there's a role that the Linux Foundation can play in being a neutral arbiter to try to get some of those projects more coalesced into a common environment that people might be more comfortable with moving from distribution to distribution?

JZ: Yeah; so first of all I think the market is a great--is one of the best mechanisms for sort of shaking out which is the best project and that's not just the market for people who purchase things but the market for people who use different Open Source projects. I mean you know there is--mutually the best projects get that critical mass and then lots of developers wanting to, Ruby is a great example of kind of coming up and building this tremendous developer community and actually taking developers from other efforts and bringing them over to their community and that's an example of--the market does that well. In Linux you have similar competition going in terms of multi-media frameworks; there's a lot of different ones out there but GStreamer seems to be one where a lot of people are coalescing and that's because it's great technology and there are interesting developers working on it. The role we play in terms of getting a consensus to happen is really in brokering forums where the developers across those various projects that make up all the components that are in a typical Linux operating system together so that they can quickly decide you know well this one it's worth going our separate ways or this one --we both pretty much think alike and we want to work together. And we do what we're calling our Linux Collaboration Summits where we bring, invite-only, 200--300 of all the key maintainers for the various components that make up Linux together in one place. They have these discussions and it just accelerates that sort of shakeout in making decisions about what--.

JT: And locked the door--two desktop enter, one desktop leaves--.

JZ: Yeah; [Laughs] exactly. It's like a cage match of who is going to do it.

JT: Right.

JZ: But it's funny; the idea for these events and for having this forum came out of the fact that these debates can go on in a mailing list for eons. You've seen these right; flame wars that never seem--everyone has got to get the last word.

JT: It's harder to call somebody, you know like a blankety-blank-blank, and when you're face-to-face with them and they react--.

JZ: Right; but there's also cloture. There's the idea that at a certain point the debate needs to end. Like we got to go, right; you know whereas an email it just tends to be this infinitum.

JT: So you mentioned a legal defense fund, so I will go there for a second.

JZ: Sure.

JT: The open running sore which is SCO seems to have resolved for the time being but there's still lots of FUD and perhaps even real issues out there about the legal standings of Linux's intellectual property.

JZ: Yeah.

JT: Where--do you see this ever having an end?

JZ: I do and here and is where we're going to go. The big issue right now is not any individual suit like the SCO suit. I think that's been put to rest. The issue today is really a matter of education amongst the legal community about how Open Source licenses work in terms of the legal community in Any Town, USA or any town in the world understanding how to deal with these licenses, how to govern them and so forth. You just don't have this wide body of understanding that you have in proprietary license legal transactions. And that's pretty much changing; you really see the legal community embracing these and at a certain point it will be like yeah we do Open Source licenses or we do proprietary license, potato--potato--it's just different.

JT: Right; but to some extent is--aren't we going to need software patents to go away before there's really a zero-risk environment for any software--not just Open Source?

JZ: Certainly I'd love to see that. I think that the patent world for software right now is just a train wreck. You know the fact that so many of these crappy patents have been issued is just a blight on the software industry. Would you rather spend the money on lots of lawsuits or would you rather spend the money on coding? I think I know what most consumers of technology would want; they'd rather spend it on cool and interesting code. But that's not how the world works today. I think there's a process to go through in reforming the patent system before we get to a world with no patents, and we obviously advocate for that world.

JT: Every year has been the year that Linux is going to make it on the desktop for at least the last four or five years.

JZ: Uh-hm.

JT: What is in your mind the thing that is most holding it back at this point?

JZ: So I should just make a prediction from now on that this isn't the year of the desktop so that I can be right for once, right. This is not the year of the desktop; nothing is going to happen this year. Let me instead of making a prediction about the year of the desktop look at what you fundamentally need to be successful in a desktop computing market besides an incumbent monopoly right because we obviously don't have that and Microsoft does, so that's a big thing. But what you need is technology that's functionally either equivalent or better to the main market incumbent and competitors. Linux just kind of got there with Ubuntu; I mean wouldn't you say that in the last year Linux has achieved the degree of sort of plug and play easy install you know cool windowing, very easy to interact with desktop functionality for the first time?

JT: I would certainly say that Ubuntu has--for me become the gold standard of are you as easy to install as--including against Windows.

JZ: Uh-hm; yeah and when Ubuntu does well everybody does well right because it's all Open Source, and so everybody uses that stuff. And so it--the bar when Ubuntu raises it everybody raises with it, so that's done. The second thing you need to be really successful in the desktop is to have good device support, right; you need to-- your video cards need to work, your wireless cards need to work, your storage and USB devices need to work well, and Linux has broken through in that area only as of recent history as well. Intel did a terrific job Open Sourcing.

JT: Graphic yeah.

JZ: Their video drivers are just really a big deal for Linux; you know you see wireless cards now being supported out of the box with Linux and obviously the USB and all those components work very well too. That's working. Application availability, lots of applications available for the platform; we're short on that but because most people now are starting to access their computing experience through a Web browser that's becoming a little less significant right; so still helping Linux.

JT: Right; and in fact I think there's an interesting thing I'm seeing there where for example if I was Sony right now I would be saying if I don't make Sound Forge available on Linux, Audacity is going to eat my breakfast because it is a multi-platform sound editor and free versus whatever Sound Forge is charging--when you don't have a Linux version all it does is encourage more development of Audacity.

JZ: Uh-hm; but now the next component that the Linux desktop needs to be successful is a critical mass within mainstream computing sales channels. And when you're talking desktop there's an enterprise desktop and there's a consumer desktop and we'll kind of look at both. In the enterprise desktop I think you have a channel and a set of requirements that are good for Linux today, right. I want a locked down desktop that's functional for my workforce that enables them to get their job done all day that can be centrally managed by a System Administrator that lowers the overall cost of managing the enterprise, better security, and so forth. And so you're starting to see companies like Peugeot in Europe wholesale switching to the Linux desktop. You're seeing educational institutions do this as well. And so there you're seeing glimmers of everything kind of coming together for the Linux desktop. In the consumer world, although I believe the raw components for a killer Linux desktop are there, it works great, it looks cool; the device support for it is there. You know applications are now through the browser so it doesn't--and the cross-platform applications to your point are really the key to success for app developers are here, so the apps are kind of there. The problem right now is that we don't really have the sales channel setup for Linux very well. I mean like most PCs are sold through retail stores. I think something like 50--60-percent of all consumer PCs are sold through retail of which Best Buy owns I think 40-percent of that channel--35, 40-percent, and you don't see a lot of Linux desktops in Best Buy. You see Best Buy and the Geek Squad offering services to remove the promo-ware for pay on Windows, right, which is just--there's got to be--there's something wrong there, so I'm hoping that Linux has a role to play--.

JT: I don't know; if they were selling Linux they'd be losing revenue stream from that.

JZ: Herein lies the problem, right; but when an interest to make money directly conflicts with what a consumer wants to experience, nobody wants to pay to remove the adware from their--it's absurd, right. It's like they do it but they don't -- there's room for innovation; there's room for somebody to come in and enter that market with a bigger and a better value-proposition so I'm hopeful that if we can figure out ways and Ubuntu again leading through getting a boxed version of their operating system on the Best Buy shelves just a couple weeks ago is an example of where there's inroads being made. But Ubuntu being, it's great to be on the box on a shelf in Best Buy, but where they really need to be is in the computer [Laughs] preinstalled and on a shelf in those types of stores.

JT: Right.

JZ: We're not there yet.

JT: It's the same deal as like Firefox. The reason there's so many IEs as out there is you have to do something to get Firefox--same deal with Linux. One of the ironies I find is that there is in fact a POSIX-based operating system out there that has significant market share that nobody talks about which is BSD because it lives inside of every Mac that's sold these days and I kind of wonder if Apple didn't get the right idea and maybe Ubuntu needs to kind of drive in that direction of just make something that doesn't look like Linux, but has the power of Linux under it. I mean--

JZ: It doesn't have to Ubuntu that does that; I mean I think you already see small startups. There's a company called gOS that's creating a very different UI around the core, 3-D Windowing functionality that you get in Ubuntu or any Linux desktop for that matter. That's very different from the experience that you get on a typical Linux desktop and much more user-friendly. There's a company called Zonbu which is building a cloud-based notebook that has a different type of UI experience from what you would get in a typical mainstream Linux desktop. There's a company called CherryPal announced a product this morning which is based--has a Debian-based device that really has the entire community experience through a Firefox browser and they feel that's the UI that is going to work and all the applications will reside in a cloud. So we still have to shake that out but all the underlying tools, frameworks are there to really create a customized and easy-to-use user interface. The big difference with Linux as opposed to Apple or Windows is, Apple really tries to create an easy-to-use platform that everyone can use; the same with Windows. With Linux you can take the tools and create a sports enthusiast operating system. You can create an operating system that's tailored to kids; you can create an operating system for women over 50. You--and still maintain application compatibility across each of those because of some of the work we do around the Linux standard base, but really tailor a computing experience to very tight demographics. I think when you are able to do that at a low-cost, work with an Asian manufacturer to build a device for you for $200 per device and combine that with this free software and tailor that to very tight markets you're going to see lots of cool things coming out. And trust me; I've seen business plans for these. I've seen prototypes of these kinds of devices. Get ready in late 2008 and most of 2009 to see lots and lots of these types of new computers coming out.

JT: Let's finish up by talking a little bit about mobile because I know that's of big interest to you and it seems to be a very hot area for Linux right now. We're seeing a lot of cellular manufacturers especially moving to Linux now; what is their big advantage do you think that Linux brings over to something like Symbian in that environment?

JZ: Well until two weeks ago I would tell you that the cost is one big advantage [Laughs] but Symbian just announced that they're Open Sourcing their platform and going to make it freely available, so that difference has gone away. I think a couple of the big differences for Linux versus Symbian or even the iPhone operating system or any other, you know RIM or any of the other competing mobile operating systems out there is one, you can take Linux and brand it your own. NEC or Panasonic or LG or Samsung, one of the larger device manufacturers can create their own branded operating system and not only build their own brand but keep a lot of that margin that derives from the software experience for themselves, and that's a huge advantage for those device manufacturers that are the primary channel for that mobile computing experience. Linux obviously from a cost-perspective keeps things low; Linux from a technical support perspective because it works on almost every CPU architecture because it supports almost every off-the-shelf component that you find in a typical mobile phone is hugely advantageous when you're building these types of devices. And that's what driving the market and that's really not replicated in any other platform out there. And then the final thing which we talked about earlier is the fact that Linux continues to innovate from unexpected areas of computing that you know benefit mobile, right. Who would have thought that you know the need for real time functionality on Wall Street would translate into better real time capability for a Linux mobile operating system? The reality is that that kind of cross-pollinization happens in Linux and it's definitely not happening on any other mobile platform.

JT: Jim Zemlin, who's the Executive Director of the Linux Foundation, thank you for talking to us, and we look forward to many great things from the Linux Foundation.

JZ: Thanks

You might also be interested in:


They mention Ubuntu having to follow the Apple standards in ease of use and beauty of the user interface. Mark Shuttleworth (the "boss" of Ubuntu) acknowledged that in several interviews, so Ubuntu "is aware" of what it should be heading to, there are scheduled UI changes for the Ubuntu from the upcoming October version, I hope they won't be canceled again like it happened in the past. Unlike other distros, Ubuntu has canceled several times any major UI changes people have been asking for, hope the October version will meet our expectations.

Popular Topics


Or, visit our complete archives.

Recommended for You

Got a Question?